It’s no secret that the hot and sweaty dog days of summer are over. There’s fresh snow on the mountain tops, grey skies, and even a few changing leaves around town. Over the last few days we’ve noticed a big dip in temperatures around the Rocky Mountain region and we’re beginning to enter a period with seasonally average temperatures, which is a nice departure from what has become the hottest year on record across the globe. However, we seem to have returned to 90° temperatures after having similar cool downs in the last two months. So, there are some real questions
I70 Eastbound Near Frisco on September 16th, 2016
as far as weather cooler temperatures will continue and, realistically, what Fall and Winter will have in store for us.
Let’s dive into these questions, but first, we’ll examine how seasonal forecasts are made, their strengths and shortcomings. As always, I’m going to break this into sections. If you want to skip the science-y stuff and go right to the forecast! No problem, just skip ahead.
- Climate and Weather Models
- The Farmer’s Almanac
- El Nino/La Nina Impacts on Colorado Weather
- Fall and Winter Forecasts
Before we continue, something important to note is that climate and weather are NOT the same thing. Weather is short term, smaller in scale and much more erratic. Climate is long term, larger in scale (think all of Earth) and follows trends.
Climate and Weather Models
Trade offs, trade offs, trade offs! It’s no secret that we are not very good at predicting the weather far in advance, but few people actually know why it’s so difficult. It’s comes down to trade offs between length of prediction (how far into the future we’re forecasting) and the size of the area we’re trying to predict. Why? Why does this trade off exist? We’re going to start with something that I saw from my professor, Dr. Sam Ng, multiple times when I was in school. Get out the bongos and get ready to snap your fingers because it’s a poem! Yes, that’s right, a poem by Lewis Fry Richardson who actually changed it from someone else’s poem, but that doesn’t matter.
WRF Model of Hurricane Sandy
“Big whirls have little whirls,
That feed on their velocity;
And little whirls have lesser whirls,
And so on to viscosity.”
Ok, I know what you’re thinking, “This poem sucks”, and you’re not wrong, it does. However, it perfectly describes our current issue with weather and climate modeling. When we talk about the atmosphere we see interactions everywhere. We see the transport of temperature, moisture, soil, gases, mass, aerosols, etc. These motions don’t exist only at large planetary scales, they shrink down to be infinitely small. They also exist in three dimensions that can interact with each other! That’s a “duh!” statement, but we have to make sure that we’re on the same page. Each parameter of the environment has its own unique set of governing equations in addition to the overall equations that describe the motion and evolution of the atmosphere and current computational power isn’t enough to get long range forecasts on small scales that can predict snowfall on a neighborhood by neighborhood basis. Hence, why there are so many botched forecasts.
For example, the NCAR supercomputers that we use to run the Weather Research and Forecasting (WRF pronounced warff) model can compute 5.34 quadrillion calculations per second and yet, they are still extremely limited in their processing abilities for detailed models.
It all comes down to grid spacing within the model structure. Our most detailed operational model, the High Resolution Rapid Refresh (HRRR), operates at 3km grid spacing and forecasts out to 24 hours in the future. It is generally assumed that, in order to resolve a weather feature, three grid points are needed. This means that our most detailed model can only resolve features that are bigger than approximately 5.5 miles. Now 5.5 miles sounds small, but it could mean the difference between the model predicting that downtown Denver will get hit and then only the areas west of Wadsworth seeing snow and Denver not getting any. Sound familiar from last year when the foothills and southern areas got snow and Denver saw almost nothing?
HRRR Model Forecast Vs. Observed Storm System- Good but not perfect
Due to its high resolution, the HRRR can only realistically be run for short forecasts in the future before it becomes computationally expensive/problematic. It is very reliable for short term forecasts and detail in small areas, but it comes at the cost of only being short term. By comparison, our climate models do very well with long range forecasting, but at the price of detail. We can accurately predict future climate patterns, but we won’t be able to have the fine details of individual weather systems and rainfall/snowfall. So, there’s that trade-off mentioned earlier. We can either use all our processing power on looking far into the future or we have to focus it on looking at very small areas. We can’t do both.
Now, this is assuming that all the data that we have pumped into these models is absolutely perfect and, you know what? It’s just not. We do a fairly good job of making measurements of the atmosphere and feeding it into the models, but the instrumentation sites are sparse and have limitations and that can create problems. Imagine an errant temperature measurement propagating forward in space and time every five minutes for thirty days! SCARY! Overall, our long range (climate) models and very short range (weather) models perform well. It’s those intermediate models that tend to have issues. So, when we are asked what the future holds and what the models are saying about seasonal weather in a particular region, it’s not as straight forward as it sounds.
The Farmer’s Almanac
For some reason, people LOVE the forecasts presented by the Farmer’s Almanac. I’m not entirely sure whether it’s due to nostalgia, good marketing, or maybe people just like to say “almanac”. It is kind of fun to say. Sounds a bit like a french pastry. “I’ll have two croissants and a raspberry almanac!” Anyway, the Farmer’s Almanac really isn’t any more reliable that any other forecast out there (maybe AccuWeather’s 90 day forecast) and its approach to forecasting weather is fairly common sense. They don’t have a gypsy fortune teller tied to a weather tower or anything.
According to their webpage-
“We derive our weather forecasts from a secret formula that was devised by the founder of this Almanac, Robert B. Thomas, in 1792. Thomas believed that weather on Earth was influenced by sunspots, which are magnetic storms on the surface of the Sun.”
Apart from the rest of their “secret formula”, which I’m fairly sure is just a fancy way of saying “climatology”, sunspots can be an indicator of weather patterns. EVERYTHING in our atmosphere is driven by solar radiation. Without it, we’re not evaporating water for clouds, temperatures would remain constant and fronts wouldn’t exist (neither would high and low pressure centers), nothing would be happening. So, when a sunspot appears and releases increased amounts of solar radiation, chances are high that we’re going to see additional warming. This is a gross oversimplification of the processes involved, but it conveys the point.
Now, there’s another possible explanation for the love of The Farmer’s Almanac and that’s what’s known as confirmation bias. Many people swear that it’s the only accurate source for weather and, after hearing that enough, others start to believe it and will look for ways (intentionally and unintentionally) to show that it is right. So, when it nails a forecast, everyone praises the Farmer’s Almanac like they did with the “Polar Vortex” outbreaks. However, when Boston had its record breaking snowfall in the 2014/2015 winter, the Almanac had failed to predict it and no one paid attention to it. A prime example of confirmation bias where you only search for data that supports what you already believe. Realistically, it is a good source of rough estimates, but it’s no better than any other source at predicting weather.
As a side note, the Polar Vortex really doesn’t travel into America. Al Roker is an idiot. The Polar Vortex is named for the region that it circles which is…. the north pole. What Al Roker was referring to was an Arctic Cold Front which brings very cold arctic air southward.
El Nino/La Nina Impacts on Colorado Weather
Higher than normal sea-surface temperatures (yellow and red) west of South America during the 1997 El Nino
This is something that needs to be touched on from time to time to remind, even myself, to be aware that there are problems with attributing Denver’s weather to ENSO (El Nino Southern Oscillation). Briefly, increased sea surface temperatures that extend westward for northern South America are classified as El Nino years. Conversely, La Nina years are times with cooler sea surface temperatures. Most often we hear people attribute big snow storms to one or the other.
“This 34 inch blizzard is awesome! I knew El Nino would come through!”
However, looking at the impacts of El Nino or La Nina on a single weather event in Denver isn’t realistic. In fact, there is almost a perfect 50/50 split of Denver’s largest blizzards between El Nino and La Nina years. That’s not to say that it doesn’t effect the weather. It effects weather patterns in a big way and can bring differences in temperature and moisture to different areas, but as far as Denver is concerned, there has never really been a correlation drawn between the ENSO cycle and our weather. This means that all those “We’re transitioning into a La Nina year so we can expect _______!” statements simply aren’t true.
Fall and Winter Forecasts
Over the last week, we’ve seen an extended period of 60 and 70 degree temperatures around the Denver metro area. While a resurgence of mid-80s temperatures will occur this weekend and into the beginning of next week, it looks like our 90 degree days are behind us as we will again see 70s late next week. The current models are showing a typical cool down over the next few weeks and seasonal high temperatures, but they’re also showing almost no precipitation at all. Unfortunately, that looks like a trend that will continue all winter.
Probabilities of higher or lower temperatures from October to December. Orange and reds are above average, blues are below average temperatures.
Probabilites of increased or decreased precipitation. Greens are above average, yellows are below average.
Right now, Denver is looking at a warmer than average winter according to the latest model runs and, unfortunately, that includes the Rocky Mountains as well. The figures above show increased temperature across most of the USA and an equal chance (EC) for above or below normal precipitation. One area of concern is that even if we get plenty of snowfall in the mountains, the warmer temperatures will ensure that it melts off faster and that our water supply is lower come summer. If the current observable trends are any indication, we could be entering into another drought stage later this winter and spring but only time will tell. So, from a climate/seasonal perspective, more of the same. Record breaking heat and drought over much of the United States.
As far as Denver is concerned, it looks like we’ll be warmer than average through all of winter based off of the latest models and information. How much snow or rain we’ll get has yet to be seen, but a wet winter isn’t looking particularly promising.
As always, if you have any questions, feel free to ask here or on Facebook and I’ll get back to you!
Thanks for reading!